Estimating efficiency of a South African bank using data envelopment analysis

ANALYSIS

BY

P.M.S. VAN HEERDEN

(M.COM RISK MANAGEMENT)

DISSERTATION SUBMITTED IN PARTIAL FULFILMENT OF THE

REQUIREMENT FOR THE DEGREE M.COM RISK MANAGEMENT AT THE

NORTH-WEST UNIVERSITY (POTCHEFSTROOM CAMPUS)

Supervisor: Prof. Gert van derWesthuizen

Potchefstroom

Prof. Gert van der Westhuizen Mr. Martin van Heerden Me. Christa van Heerden Mr. Lambert Taute

Thank you for all your unlimited support and help over the past year. Special thanks to Prof. Gert for all your sacrifices, advice, help and guidance. Prof. Gert, words cannot express how thankful I am for having you as my supervisor. Thank you!

I would also like to praise God for giving me the talents to write this paper. Heb 13:5-6 ~ Let your conversation be without covetousness, and be content with such things as ye have for he hath said, I will never leave thee, nor forsake thee. So that we may boldly say, The Lord is my helper, and I will not fear what man shall do unto me.

Finally, I would like to thank my father for his devotion, his life's sacrifice and love he gave to our family. Thank you for the privilege to study at the North-West University. 'Dad, this one is for you!'

adjusted performance measures, bank efficiency; scale efficiency; scope efficiency; X-efficiency; cost X-efficiency; standard profit X-efficiency; alternative profit X-efficiency; intermediation approach, production approach, asset approach, user-cost approach; value added approach; Data Envelopment Analysis; input-orientated; output-orientated; constant returns to scale; variable returns to scale; efficiency scores

The greater competition and concentration in South Africa's financial sector has put South African banks under more constraints and led to questioning of their present performance. With a greater demand for financial services and more complains about the low quality of financial services and charges being too high, there has been increasing debate about how efficient South African banks really are.

This study discusses performance evaluation, the traditional financial and non-financial measures used, and their limitations. The concept of bank efficiency is also briefly discussed, including scale efficiency, scope efficiency, X-efficiency, cost efficiency, standard profit efficiency, alternative profit efficiency and the risk component of bank efficiency.

Data Envelopment Analysis (DEA) was chosen as the most appropriate method to estimate the scale efficiency and technical efficiency of 37 districts (and 10 provinces) of one of the largest banks in South Africa. 'DEA involves solving linear programming problems that generate a non-parametric, piecewise linear convex frontier that envelops the input and output data relative to which cost is minimized' (Fare et al., 1985b:193). The intermediation approach was used incorporating both the input- and output-orientated approach under variable returns to scale.

The analyses indicated that 19 districts out of the 37 districts were not at least once fully technically efficient during the 22 months (input- and output-orientated). The same results were

maatstawwe; risiko-aangepaste maatstawwe vir werksverrigtinge, bankdoeltreffendheid; skaaldoeltreffendheid; geleentheidsdoeltreffendheid; X-doeltreffendheid; koste-doeltreffendheid; standaard winsdoeltreffendheid; alternatiewe winsdoeltreffendheid; intermediere benadering, produksie benadering, bate benadering, gebruikerskoste benadering; toegevoegde waarde benadering; Data Envelopment Analise; inset-orienteringsbenadering; uitset-orieterings-benadering; konstante skaalopbrengs; veranderde skaalopbrengs; telling vir doeltreffendheid

Die toenemende kompetisie en markkonsentrasie het gelei tot groter druk op die Suid-Afrikaanse finansiele sektor. Dit het daartoe gelei dat markoritleders die huidige Suid-Suid-Afrikaanse banke se doeltreffendheid bevraagteken het. Die groter vraag na finansiele dienste, meer klagtes oor die swak kwaliteit van finansiele dienste en die hoer koste van dienste het gelei tot "n toenemende resensering van Suid-Afrikaanse banke.

Die studie bespreek die evaluering van die bank se werksverrigtinge, die tradisionele finansiele en nie-finansiele maatstawwe wat gebruik is en hul beperkings. Die konsep van bankdoetreffendheid is ook kortliks bespreek, dit sluit in skaaldoeltreffendheid, geleentheidsdoeltreffendheid, X-doeltreffendheid, koste-doeltreffendheid, standard winsdoeltreffendheid, alternatiewe winsdoeltreffendheid en die risiko komponent van bankdoeltreffendheid.

Die Data Envelopment Analise (DEA) was gekies as die mees geskikte metode om die skaaldoeltreffendheid en tegniese doeltreffendheid van 37 distrikte (en 10 provinsies) van een van die grootste banke in Suid-Afrika te meet. Die DEA behels die berekening van liniere programmeringsprobleme wat 'n nie-parametriese, konvekse grens genereer, wat alle insette en uitsette insluit om kostes the minimeer (Fare et a/., 1985b: 193). Die intermediere benadering

Die analises toon dat 19 distrikte uit die 37 distrikte was nie een keer gedurende die 22 maande volkome tegnies doeltreffend nie (inset- en uitset-georienteerde benadering). Dieselfde resultate is bevind met die toets vir skaaldoeltreffendheid oor die 22 maande. 17 distrikte uit die 37 distrikte was nie een keer volkome skaaldoeltreffend nie (inset-georienteerd) en 19 distrikte uit die 37 distrikte was nie een keer volkome skaaldoeltreffend nie (uitset-georienteerd). Sinergie was gevind in 6 van die 10 provinsies (inset- en uitset-georienteerde benadering).

List of Tables xi

CHAPTER 1: INTRODUCTION

1.1. Introduction 1 1.2. Problem statementand motivation 1

1.3. Goal 5 1.4. Research methodology 5 1.5. Outline of study 1.5.1. Chapter 2 6 1.5.2. Chapter 3 6 1.5.3. Chapter 4 6 1.5.4. Chapters 7 1.5.5. Chapter 6 7 CHAPTER 2: BANK PERFORMANCE EVOLUATION

2.1. Introduction 8 2.2. Background 8 2.3. Bank performance 10

2.4. The role of performance measures in an organization 16

2.5. Financial measures 17 2.5.1. Profitability measurement 17

2.5.1.1. Return On Assets (ROA) 19 2.5.1.2. Return On Equity (ROE) 20 2.5.1.3. Other profitability measurements 23

2.5.2. Other financial measurements 24 2.5.2.1. Liquidity ratios 24 2.5.2.2. Leverage ratios 24 2.5.2.3. Profitability ratios 25 2.5.2.4. Efficiency ratios 25

2.6. Non-financial measures 30 2.7. The Balanced Scorecard (BSC) 31

2.7.1. Financial perspective 32 2.7.2. Customer perspective 32

2.7.4. Learning and growth measures 32

2.8. Risk-adjusted measures 34 2.9. The meaning of risk-adjusted measures 36

2.9.1. Risk-adjusted performance measures (RAPM) 40

2.9.2. The concept of Value at Risk (VaR) 43

2.9.3. Economic Value Added 44 2.10. Factors influencing bank performance evaluation 47

2.10.1. Lack of understanding 47 2.10.2. Communication 47 2.10.3. Commitment 47 2.10.4. Participation 48 2.10.5. Trust 48 2.10.6. Managerial support 48

2.10.7. Social psychological factors 48 2.11. The unresolved problems of bank performance evaluation 48

2.11.1. The 'adding-up' problem of bank capital 49 2.11.2. The differences between market-based capital allocation

and actual capital 49 2.11.3. Valuing product and customer relationships 50

2.12. Summary 50

CHAPTER 3: BANK EFFICIENCY

3.1. Introduction 52 3.2. Background 52 3.3. What does efficiency mean? 54

3.3.1. Scale efficiency 55 3.3.2. Scope efficiency 56 3.3.3. X-efficiency 57 3.3.3.1. Technical efficiency 57 3.3.3.2. Allocative efficiency 58 3.3.4. Cost efficiency 61 3.3.5. Standard profit efficiency 63

3.3.6. Alternative profit efficiency 64 3.3.7. The risk component of efficiency 66

3.4.1. Inputs and Outputs 70 3.4.1.1. The production approach 72

3.4.1.2. The intermediation approach 73

3.4.1.3. The asset approach 77 3.4.1.4. The profit approach or user-cost approach 79

3.4.1.5. The risk management approach 79 3.4.1.6. The value added approach 80 3.5. Problems associated with the measurement of bank efficiency 81

3.6. Summary 82

CHAPTER 4: DATA ENVELOPMENT ANALYSIS

4.1. Introduction 84 4.2. Background 85 4.3. The average costfunction 86

4.4. Data Envelopment Analysis (DEA) 88

4.4.1. Input-orientated 92 4.4.2. Output-orientated 93

4.4.3. Slacks 94

4.4.5. Variable returns to scale (VRS) 98 4.4.6. Cost minimization and profit maximization 101

4.4.7. Environmental factors 103 4.4.7.1. Method 1 104 4.4.7.2. Method 2 104 4.4.7.3. Method 3 105 4.4.7.4. Method 4 107 4.4.8. Congestion 109 4.4.9. The Malmquist index 112

4.4.10. Advantages and disadvantages of DEA 115

4.5. Stochastic Frontier Analysis (SFA) 117

4.6. The thick frontier 117 4.7. Efficiency scores 118

5.1. Introduction 121 5.2. The intermediation approach 121

5.3. Efficiency estimates 123

5.4. Summary 147

CHAPTER 6: CONCLUSSIONS AND RECOMMENDATIONS

6.1. Introduction 150 6.2. Conclusion 151

Figure 2.1: The framework of bank performance 12 Figure 2.2: Policies to maximize the bank's equity value 14

Figure 2.3: Framework of risk-return bank performance 36

Figure 2.4: Example of the concept of VaR 44 Figure 3.1: The dual role of production efficiency and intermediation

efficiency 59 Figure 3.2: Allocative efficiency and technical efficiency 60

Figure 3.3: Cost efficiency 62 Figure 4.1: The production frontier 85

Figure 4.2: A DEA model showing an efficiency frontier 91 Figure 4.3: Illustration of an input-orientated organization 93 Figure 4.4: Illustration of an output-orientated organization 94

Figure 4.5: Illustration of an input slack 94 Figure 4.6: Illustration of an output slack 95 Figure 4.7: Illustration of a production possibility set 97

Figure 4.8: Calculating scale economies in DEA 99 Figure 4.9: The differences between input- and output-orientated

technical efficiency measures and return to scale 101 Figure 4.10: Illustration of an isoquant reflecting input congestion 109 Figure 4.11: Illustration of input congestion and efficiency

measurement 111 Figure 4.12: Illustrating the catch-up and boundary-shift factor 113

expense resulting from a 25% increase in volumes in the branch-based retail business of a money central

bank 19 Table 5.1: Efficiency estimates for district 2 125

Table 5.2: Efficiency estimates for district 4 127 Table 5.3: Efficiency estimates for district 12 129 Table 5.4: Efficiency estimates for district 13 131 Table 5.5: Efficiency estimates for district 25 133 Table 5.6: Efficiency estimates for district 26 135 Table 5.7: Efficiency estimates for district 31 137 Table 5.8: Efficiency estimates for district 34 139

Table 5.9: Districts and provinces 140 Table 5.10: Efficiency estimates for province 2 142

Table 5.11: Efficiency estimates for province 6 144 Table 5.12: Efficiency estimates for province 9 146

Chapter 1: Introduction

1.1. Introduction

Banks fulfill the primary need for financial services in the South African economy. Customers need to believe that the financial services they receive provide reasonable value for money. Various increases in bank fees over the last decade or two have led to questioning of the efficiency levels of South African banks. Are the higher bank fees the result of a lack of efficiency within the banking sector?

1.2. Problem statement and motivation

Until the late 1980s the South African financial sector was dominated by five commercial banks, namely Standard Bank, First National Bank, Volkskas Bank, Nedbank and Trust Bank. During the 1990s the banking sector underwent re-organization and consolidation, where Volkskas Bank, Allied Bank, United Bank and Sage Bank merged to create the Amalgamated Banks of South Africa (ABSA) (Akinboade & Makina, 2006:107). Today the four largest banks in South Africa are First National Bank, Nedbank, Standard Bank and ABSA (Hawkins, 2004:183).

According to Mboweni (2004:1) South Africa has established a well-developed banking system over the past decade. South African banks are also well utilized and managed in sophisticated risk-management systems and corporate-governance structures. With the required systems available, why should South African banks not be efficient? Akinboade and Makina (2006:107) stated that the reorganization of the banking sector during the 1990s led to the establishment of banking services to poor communities, which were neglected during the apartheid era.

From the end of apartheid an increasing number of black people have entered the formal economy, demanding more banking and other financial services in townships (Okeahalam, 2006:105). Finscope (2007:1) stated that the number of banked South Africans has increased

from 46.6% in 2005 to 5 1 % in 2006. Consumers however complain that the present service quality is low and that bank charges are too high (Okeahalam, 2006:105). Does this mean the increasing need for financial services has impeded the banks' ability to be efficient?

According to SARB (2007:17) the South African banking sector is highly concentrated. The Herfindahl-Hirschman Index (H-index) can be used to measure the concentration in the banking industry. A H-index below 0.1 indicates that there is no concentration and a H-index above 0.18 indicates a high level of concentration. A H-index between 0.1 and 0.18 is an indication of moderate concentration (Bank Supervision, 2007:2). The H-index showed an estimate of 0.184 during December 2006 (see Figure 1.1), indicating the great dominance by the four largest South African banks (Bank Supervision, 2007:2). Figure 1.1 also shows how the concentration of the South African banking industry increased from an estimate of 0.131 in 2001 to 0.184 in 2005 and 2006.

Figure 1.1: The H-index for the South African banking system (2001-2006) Index

0„1S I 1

20Q1 2002 2003 2004 2005 2006 Source: Bank Supervision (2007:2).

In 1994 the four largest banks owned about 83.8% of total assets owned by the banking sector and about 87.4% in 2004 (Mboweni, 2004:1). According to Mboweni (2003:9) the four largest South African banks had about 83% of total deposits during 2003. Okeahalam (2006:105) also stated that these four banks are controlling over 85% of total deposits and assets in South Africa. These four banks also handle about 85.17% (March 2005) of the banking business in South Africa (Van der Westhuizen, 2006:1). These figures are another indication of the great market power these four banks have. The increased interest margin banks experienced towards the end of 2000 is also an indication of the strong market power present in the South African financial sector (Bank Supervision, 2002:13). At the end of March 2005 the market shares for ABSA was 25.06%, for Nedbank 23.92%, for Standard Bank 30.17% and for First National Bank 20.85%.

The South African banking industry is more concentrated that the British banking industry and is less subject to international competition. That is why the South African Reserve Bank (SARB) (2000:170) claimed that the findings of the British banking industry investigation, which was done by the Cruickshank Commission, are also applicable to the South African banking industry. The Cruickshank Commission came to the following conclusions (SARB, 2000:170):

• The banks were making monopolistic profits from the payment system. • The banks were allowed to write their own rules.

• The banks were not supplying sufficient useful information to consumers.

This is another indication of the dominance the four largest banks may currently have. The question now is, are these four banks using their dominance to increase bank efficiency? This leads to another question: Are bank customers paying too high bank fees for the financial services they receive, or are the four large banks using their dominance to become more

profitable rather than being more cost efficient?

With compliance practices and regulated accounting changing, financial costs are increasing in the bank industry (Hawkins, 2004:196). Operating costs have also outgrown bank incomes during 2000 (Hawkins, 2004:196). Akinboade and Makina (2006:117) stated that the ratio of deposits to Gross Domestic Product (GDP) maintained an upward trend during the period of 1994 to 2002, averaging 61%. This suggested that local short term savings have been more effectively mobilized. However, according to Hawkins (2004:200) savings accounts are costly, accruing fees for both withdrawals and deposits. Banking Supervision (2002:50) stated that staff costs continued to rise, while employment and expenses associated with branch closure are declining. It is also extremely difficult to determine if consumers are receiving financial products and services at a fair price. Hawkins (2004:197) stated that South African banks compete by advertising interest rates, while charges and fees are rarely revealed (Hawkins, 2004:197). The increase in non-interest income to about 50% of banks' income suggests that customers may not be paying fair prices for financial services (Hawkins, 2004:197).

Apart from the competitiveness in the South Africa banking industry, competitive constraints removed from this industry (SARB, 2000:160), are leading to challenges the South African banks must overcome. These challenges include the following (SARB, 2000:170):

• Ensuring that the financial sector remains systematically stable in a rapidly changing technological environment and sharply increased competitive conditions.

• Ensuring that banking services are delivered to the whole community, including the poor.

Another challenge according to Arora and Leach (2005:1726) may also be the cost of providing financial services on a small scale.

Greenwood and Jovanovic (1990:1076) and Akinboade and Makina (2006:103) stated that financial intermediaries have the ability to allocate resources efficiently and the ability to promote long-run economic growth. According to Hawkins (2004:196) there is still room for the South African banks to improve their efficiency. Thus the primary motivation for this study is to shed light on the present standing of the efficiency levels of a large South African bank.

1.3. Goal

The lack of data about each individual bank branch led to the use of districts. The goal of this study is to estimate the efficiency of 37 districts of one of the largest banks in South Africa covering a 22 month period. These 37 districts are also aggregated into 10 provinces to determine if synergies are present in the bank structure.

1.4. Research methodology

To establish the required background a literature study on various topics will be done. This research will enable one to perform an empirical study, which involves the estimating of the efficiency levels of one of the largest banks in South Africa. DEA will be used to estimate the relative efficiency of the 37 districts.

The DEAP (version 2.1) program developed by Coelli (1998) will be used. The intermediation approach is best suited for the available data and both the input- and output-orientated approach under the variable returns to scale approach will be used. Variable returns to scale has fewer restrictions than constant returns to scale which will not function under conditions such as imperfect competition and constraints on finance (Coelli et a/., 1998:150). Both technical efficiency and scale efficiency of the 37 districts will be estimated.

1.5. Outline of study

A summary of each chapter's contents will follow to indicate what the reader can expect from each chapter.

1.5.1. Chapter 2

In chapter 2 bank performance evaluation is discussed. The meaning, role and development of a performance model in the organization is discussed. Traditional performance measures, including financial and non-financial measures are discussed with their advantages and disadvantages. The risk factor and why bank performance evaluation must be accompanied by the risk factor is also discussed. The factors influencing bank performance evaluation and the unresolved problems in bank performance evaluation are also discussed.

1.5.2. Chapter 3

In this chapter the meaning of efficiency is discussed including the different concepts of bank efficiency. This includes scale efficiency, scope efficiency, X-efficiency, cost efficiency, standard profit efficiency, alternative profit efficiency and the risk component of bank efficiency. The steps in measuring bank efficiency are also discussed and this includes the approaches available for choosing the appropriate inputs and outputs. The problems associated with the measurement of bank efficiency are also discussed.

1.5.3. Chapter 4

Data Envelopment Analysis (DEA), as the chosen technique in this study, will be discussed. The DEA involves the use of linear programming methods to construct a non-parametric, piece-wise frontier across the data (Coelli et al., 1998:140). The following factors, that influence the construction of the DEA model, are discussed. These factors are: input- or output orientation; slacks; return to scale properties; cost minimization or profit maximization; environmental factors and congestion. The Malmquist index as well as the advantages and disadvantages of the DEA

are also discussed. Other efficiency measurements, namely, the Stochastic Frontier Analysis (SFA) and the thick frontier are also briefly discussed. Efficiency scores that are generated from different efficiency measurements are also discussed.

1.5.4. Chapter 5

This chapter involves the efficiency estimates generated by the DEA analysis. Both technical efficiency and scale efficiency of 37 districts over a 22 month period will be estimated and it includes both the input- and output-orientated approach under variable returns to scale. The intermediation approach will be used to classify the inputs and outputs used in the DEA analysis. The presence of synergies will also be evaluated to determine whether the 37 districts generate a greater efficiency estimate as a province (group) rather than as a district.

1.5.5. Chapter 6

This chapter contains the findings and concluding remarks which the literature study and DEA analyses revealed.

Chapter 2: Bank Performance Evaluation

2.1. Introduction

The goal of this chapter is to introduce bank performance evaluation. In doing so, attention will be paid to aspects like why performance measurements are important (section 2.3); the role of performance measures (section 2.4) and which factors keep the bank from achieving its goals of maximizing bank equity (section 2.3). How the main performance indicators are chosen (section 2.3, 2.4); which flaws/drawbacks and limitations do performance indicators have (section 2.3) and the different types of performance measures (both financial and non-financial measures) (section 2.5, 2.6) will be discussed. Also in this chapter the weaknesses of financial measures, non-financial measures and of the balanced scorecard will be discussed (section 2.5, 2.6, 2.7). How some of these weaknesses can be overcome by introducing risk-adjusted measures (section 2.8); the meaning of risk-adjusted measures, their weaknesses and the different types of risk-adjusted measures (section 2.9); the factors influencing bank performance evaluation (section 2.10) and the unresolved problems (section 2.11) are discussed.

2.2. Background

Greater competition in domestic and international markets and fundamental structural changes are forcing organizations to be more flexible and productive, with greater concentration on serving the customer (Hartle, 1997:45). Kimball (1997:24) stated that the greater specialization and focus in commercial banks were permitted by the 30 years of evolution in bank structures and management. During the 1980s organizations became more performance-orientated, which led to an era of 'management by objectives' (Hartle, 1997:46). However, until 1994 South Africa's banking sector had been deteriorating because of political isolation and restrictions on bank branching (Mboweni, 2004:1). The restrictions on bank branching in the U.S.A. caused limited bank performance and vulnerability to localized economic distress.

However, these legal restrictions to interstate branching was removed in June 1997 (Neely & Wheelock, 1997:24).

Kimball (1997:24) stated that the economy realized that there was an increasing need for specialized expertise in banking, for example, expertise in cash management, international banking and asset-based lending. Banks began therefore to specialize in different product lines and this was soon designated as a line of business (Kimball, 1997:24). Banks began to organize and manage themselves as a collection of different lines of businesses, each with a different product, customer, distribution channel or geographic mandate operating semi-autonomously. This gave rise to new issues concerning performance measurement, risk management, resource allocation and it resulted in the strategic use of branches (Kimball, 1997:24).

Hartle (1997:65) stated that effective performance management has some of the following phases, namely, planning, managing and reviewing performance. Kimball (1997:24) stated that top management needed profitability reports for each branch to measure performance. According to Hartle (1997:66) performance review gives the organization the opportunity to step back from the daily activities, helping to analyze performance trends and to plan for the future. Verma (1992:279) stated that performance measurement contributes to management control, where it evaluates whether results from planned actions were realized. However, this caused problems because branches often shared customers, products and distribution channels. Kimball (1997:24) claimed that the answer to this problem was to gather more information than just organizational profitability. According to Verma (1992:279) traditional measures of profit were no longer appropriate for the motivation of desired behaviours such as greater productivity and focus on customers. Organizations were therefore pressured to redesign their performance systems. Rather than the traditional measures like financial profit and cost, new categories were

included for measuring productivity, quality and customer satisfaction (Berliner & Brimson, 1988:62).

Each line of business generated different exposures to interest rate risk, credit risk, prepayment risk and operating risk that can expose the bank to possible losses. Top management realized that reporting systems were needed to monitor financial results and potential risks (Kimball, 1997:25). The profitability of branches varied directly with the riskiness of their portfolio and operations (Koch & MacDonald, 2003:118), making resource allocation based only on gross profitability not sufficient. Top management became therefore aware of the need to measure returns on a risk-adjusted basis (Kimball, 1997:25). New performance reporting, risk management systems and new data bases were needed. New analytical approaches permitting top management to objectively weigh costs, benefits and risks were also needed (Kimball, 1997:25).

2.3. Bank performance

Mester (2003:3) stated that for a bank to achieve outstanding performance it means that the bank succeeded in maximizing the shareholders' wealth, in other words, maximizing the market value of a firm's common stock. Company performance should be measured against an objective. Without an objective the company will have no criterion on choosing investment strategies and projects (Mester, 2003:3). Performance measurement is important in keeping the company on track in achieving their objectives (Armstrong, 2000:4). "The main purpose of performance measurement is to align the goals of individual employees and the bank as a whole" (McDonell & Rubin, 1991:56). Mester (2003:3) stated that efficiency is a measure of the deviation between actual performance and desired performance. The objective of performance evaluation and the information used depends on the perspective of the evaluator (Gardner & Mills, 1994:667). The forefront of managing company performance is to measure the

competitive productivity strategies, quality improvements and speed of services (Johnson, 1996:846-848).

Fitzergerald et al. (1993:6-8, 19-22) stated that there are no boundaries in selecting main performance indicators, because they are used for random purposes. Fitzergerald et al. (1993:6-8, 19-22) continue stating that the main performance indicators include a combination of financial, market/customer, competitor, human resource, internal business process and environmental indicators. It is important to combine both cost efficiency estimates with profitability tests so as to evaluate financial firm efficiency. One needs to evaluate a bank's ability to use resources effectively in producing products and services (cost efficiency) and their skill at generating income from these services (profit efficiency) (Spong et al. 1995:5).

Avkiran (1997:232) stated that the main steps that should be followed in developing a performance model are as follows:

• Review the corporate objectives and strategies.

• Define branch performance in terms consistent with corporate objectives. Identify those performance variables that are critical to the bank's success. Reflect the mix of personal banking business at branches.

• Identify the potential variables that are associated with branch performance. Identify those potential variables controllable by management.

• Determine the sources of data for performance and potential variables. • Develop multivariate measures for data collection.

• Analyze data through multiple regression and other techniques. Determine the sets of potential variables explaining each of the performance variables.

• Examine the results of the analysis and raising performance through reconfiguring branches.

Figure 2.1: The framework of bank performance

Market price of bank's shares

Return/risk trade-off Shareholders' preferences Executive decision making Business banking performance Personal banking performance

I

The catchment-area-specific, non-controllable potential variables capture market information, socio-economic, demographic and market information. These variables interact with branch-specific controllable potential variables, resulting in branch performance outcomes (Avkiran, 1997:225). These outcomes define the personal banking performance. The executive decision making is influenced by the nature of the economic and regulatory environment. It is also influenced by the shareholders' preferences and feedback on performances of business banking, personal banking and operations (Avkiran, 1997:226). Executive decisions involve the trade-offs between maximizing returns and minimizing risks. Capital markets therefore assess the overall performance of the bank. Keeping in mind the bank's managerial actions and the accurate pricing of shares (Avkiran, 1997:226).

In order to maximize a bank's equity Hempel and Yawitz (1977:25-27) identified some key variables. According to them these variables include gross receipts from assets (interest income and non-interest income from loans), cost of liabilities (interest expense on deposits), overhead costs (non-interest expense), taxes and risk premiums, which are added to the risk-free interest rate. Managerial decisions as one of the key variables must also be examined, because they produce four key activities influencing the equity value. These key activities include spread management, control of overhead, liquidity management and capital management (Avkiran, 2006:277). Spread management is the management of the difference between gross revenues (interest income plus non-interest income) and interest expenses, where the desirable state is to sustain a high positive spread over time (Avkiran, 2006:278).

Spread management also includes the control of non-interest expenses, where the control of overhead costs or non-interest expenses is the ability to minimize these expenses while maintaining a high spread (Avkiran, 2006:278). Liquidity management is the ability to convert short-term assets into cash to meet unexpected deposit withdrawals or funding needs and liquidity requirements of the country's central bank (Avkiran, 2006:278). Swank (1996:176) stated that liquidity management today involves borrowing from interbank markets serviced by wholesale banking and securitization of assets. Avkiran (2006:278) stated that capital management is the balancing of the level of capital in such a manner to sustain growth of assets and liabilities without decreasing public confidence or profitability.

Sinkey (1992:70-71) also developed a framework (Figure 2.2) to maximize the equity value of a bank, which stated that the objective of maximizing the shareholders' wealth is being determined by the owners' preferences, management's attitudes and decisions and the society. Sinkey (1992:71) also stated that there are six policy strategies to take into account in order to achieve maximum equity value. These policies depend on the riskiness of a bank's balance sheet and include spread management, control of the burden, liquidity management, tax

management and management of off-balance sheet activities. Tax management is the minimization of tax liabilities. The management of off-balance sheet activities, like letters of credit, lines of credit and loan commitments, is to increase fee income and to reduce the burden. The burden is the difference between non-interest income and non-interest expenses (Avkiran, 2006:278).

Figure 2.2: Policies to maximize the bank's equity value

Owners' Preferences

Bank's Primary ^ _ Society: The Objective Regulatory and

'Maximize equity Economic value' Environment

Policy Strategies to Achieve Bank's Primary Objectives

1. Spread Management 2. Control of 'Burden' 3. Liquidity Management 4. Capital Management 5. Tax Management 6. Management of Off-Balance Sheet Activities Source: Sinkey (1992:70).

Virtually all performance evaluators use accounting and other data to help in calculating the financial condition of an institution and the level of management (Gardner & Mills, 1994:667). Financial measures alone have serious limitations, because of their backward-looking nature, their limited ability to measure operational performance and their tendency to focus on the short-term (Kaplan & Norton, 2001 a:3). Backward-looking nature means that it is limited by historical financial data. The value of information in measuring performance is addressed by the 'informativeness principle' (Holmstrom 1979:83-87; Shavell 1979:56-59). This principle states

Management's Attitudes and

that the more reliable the information, the more weight it should be given with the goal of reducing the error term in a model (Milgrom and Roberts 1992:219).

Koch and MacDonald (2003:170) stated that the traditional bank performance analysis carried three basic flaws:

• It ignored the wide diversity in strategies pursued by different institutions.

• A bank's total assets no longer served as a meaningful yardstick when banks engaged in off-balance sheet activities.

• The analysis provided no direct information concerning how or which of the bank's activities contributed to the creation of shareholder value. It therefore ignores other performance benchmarks that consumer-focused managers must have considered to identify the best strategies going forward.

De Young (1997:21) stated that the accounting-based cost ratio is the traditional tool used by bank analysts to measure cost efficiency, which is difficult to interpret. For example, the banking industry has become inefficient over time, spending over 20% more on labour, materials and physical plant. These data are misleading, because this cost ratio does not control for increases in fee-based activities. It alters the relationship between non-interest expenses and assets in banks. In other words, expense ratios mislead trend analysis if product mix changes over time and in cross-sectional analysis if the banks being compared have dissimilar product mixes (De Young, 1997:21). Brown and Mitchell (1993:729-732,735-736) stated that the selection of performance measurement indicators should be:

• Driven from strategies, providing a linkage between strategic plans and business unit actions.

• Supportive of the company's multidimensional environment.

• Based on an understanding of the cost behaviour and cost relationships.

The next section describes the contribution performance measures make in the organization.

2.4. The role of performance measures in an organization

A successful organization depends upon the decision-making ability of its managers, which depends upon the availability of useable information (Milgrom & Roberts, 1992:544). Performance measurement is therefore important because it helps the organization to stay on track in achieving its objectives (Armstrong, 2000:4). It also serves as a monitoring mechanism employed by the owners of a company (Baker & Wruck, 1989:167,189).

Apart from the requirements stated by Brown and Mitchell (1993: 729-732,735-736) in the previous section, Kimball (1997:25, 36-40) argued that a good performance measurement system should be:

• Supportive and consistent with an organization's people/culture, goals, actions and key success factors.

• Driven by the customer.

• Appropriate to the external and internal environment. • Developed by a combined bottom-up and top-down effort. • Integrated and communicated throughout the organization.

• Focused more on managing resources and inputs, not just on costs. • Committed to providing action-orientated feedback.

How to measure performance still depends on each individual organization, because of different situations and circumstances (Berry et al., 2005:94). Performance is defined by individual organizations according to their business objectives and strategies (Fitzergerald et al., 1993:10). The main performance indicators are usually a combination of financial and non-financial indicators (Kaplan & Norton, 2001 a:3).

2.5. Financial measures

The American Accounting Association (1966:1) defined financial measures as a 'process of identifying, measuring and communicating economic information to permit informed judgment and decisions by users of the information'. Financial measures consist mainly of profitability measures and other financial measures that will be discussed in the following sections.

2.5.1. Profitability measurement

Although the fund transfer pricing systems helped in identifying and managing bank exposures to interest rate risk, they were not sufficient in calculating profitability (Kimball, 1997:31). From the early 1980s new cost accounting methodologies called activity-based accounting were introduced. It permitted banks to better understand the forces driving their costs and to allocate these costs to their sources (Kimball, 1997:32). However, the problem with this approach was that it did not provide enough information about the relationship between revenues and expenses. The purpose of activity-based accounting was to build cost allocation systems around business processes. This gave a better view of the relationship between transaction volumes and incremental costs (Kimball, 1997:32). Activity-based accounting made it possible to reduce the shared costs treated as overheads and therefore allocate such costs to the products or customers. However, under the activity-based accounting system the product upgrades cost would be allocated to the requesting branch (Kimball, 1997:33).

Banks began to built systems where they were able to calculate profitability by line of business and also to measure profitability of customers, products and distribution channels (Kimball, 1997:33). Insight was therefore gained concerning the differences among customers in their profitability to the bank. A study was conducted on the middle-market corporate customers of a Texas bank, where they were grouped according to their profitability to the bank (Kimball, 1997:33). The results indicated that only 30% of the customers were profitable. These profits were also twice the size of the profits of the whole business unit and this study gave banks the insight to realign service levels and to price products better. Banks also needed to analyze customers to understand the sources of these differences in profitability (Kimball, 1997:33).

Kimball (1997:34) stated that revenues are positively correlated with the sum of deposits and loans. This makes profitability very sensitive to changes in branches' deposits and loans. These insights help banks to undertake acquisitions, where acquisitions permitted banks to consolidate branches with overlapping service areas and to obtain economies of scale at the branch level (Kimball, 1997:34). The question is how will a bank determine the increase in non-interest expenses resulting from the acquistion? Using table 2.1, this problem can be addressed for the branch-based retail business of a money center bank. In table 2.1 the percentage increase in non-interest expense associated with a 25% increase in volumes will differ by activity. However, this increase is less than the increase in volumes (Humphrey, 1985:767-770). The overall non-interest expense will increase by 10%. Table 2.1 also explains the decision by banks to expand through acquisitions. However, 'if the acquired bank had overlapping distribution systems and the same cost structure as the acquiring bank then table 2.1 would indicate that 60% of the non-interest expense of the acquired bank could be eliminated after a merger'. These kind of analyses can help with the setting of goals by banks to help reduce expenses from mergers and consolidations (Kimball, 1997:35).

Table 2.1: Volume relationship percent increase in non-interest expense resulting from a 25% increase in volumes in the branch-based retail business of a money center bank

Function A B C D E Marketing 7 7 Sales 11 11 2 — 7 Transportations Processing 18 15 2 — 15 Account Maintenance — 13 3 — 11 Customer Service 11 5 2 — 10 Support/Management 6 1 ~ 4 5 Total 12 12 2 4 10 A - Branches B - Operations C - Systems D - Support/Overhead E - Total

Source: Gemini Consulting (1999: Appendix).

Most bank managers however quote Return On Equity (ROE) or Return On Assets (ROA) or other growth rates when asked about the performance of the bank (Avkiran, 1997:224).

2.5.1.1. Return On Assets (ROA)

ROA is a profitability measurement and an overall measure of bank performance, from an accounting perspective. It measures the bank's net income per dollar of assets (Sinkey, 1992:43). ROA gives an idea as to how efficient management is at using its assets to generate earnings. ROA can be calculated by dividing a company's annual earnings by its total assets. Sometimes this is referred to as 'return on investment' (Anon., 2006a:1):

R Q A =

Net _ Income

Total Assets

Some investors add interest expense to net income when performing this calculation, because they would like to use operating returns before cost of borrowing. ROA tells one what earnings were generated from invested capital (assets) (Anon., 2006a: 1). ROA can vary substantially and will be highly dependent on the industry. For this reason it is best to compare it to the ROA of a similar company. The ROA figure gives investors an idea of how effectively the company is converting the money it has to invest into net income (Anon., 2006a: 1). The higher the ROA the better it is, because the company is earning more money on less investment. For example, if one company has a net income of $2 million and total assets of $6 million, its ROA is 33%. If another company earns the same amount of income but has total assets of $8 million, it has an ROA of only 25%. Based on this example, the first company is better at converting its investment into profit (Anon., 2006a:1).

ROA provides no direct information concerning how or which of the bank's activities contribute to the creation of shareholder value. It ignores other performance benchmarks that customer-focused managers must consider to identify the best strategies for the future (Koch & MacDonald, 2003:170).

2.5.1.2. Return On Equity (ROE)

ROE is another profitability measurement and measures accounting profitability from the shareholder's perspective (Sinkey, 1992:43). It measures a firm's efficiency at generating profits from every dollar of assets. It also shows how well a company uses investments to generate earnings growth (Anon., 2006b:1). ROE is equal to a fiscal year's net income (after preferred stock dividends, before common stock dividends) divided by total equity (excluding preferred shares) (Anon., 2006b: 1):

ROE = Netjncome ^

Average _ Stockholders _ Equity

ROE is best used when it is compared to companies in the same industry. High ROE yields no immediate benefits. Since stock prices are most strongly determined by earnings per share (EPS), investors will be paying twice as much (in Price/Book terms) for a 20% ROE company as for a 10% ROE company (Anon., 2006b: 1). The benefit comes from the earnings reinvested in the company at a high ROE rate, which in turn gives the company a high growth rate. ROE is irrelevant if the earnings are not reinvested. ROE is calculated from the company's perspective, on the company as a whole (Anon, 2006b: 1).

The du Pont analysis is a way to break ROE down into three components, namely net margin, asset turnover and financial leverage. Splitting ROE into three parts makes it easier to understand changes in ROE overtime (Anon., 2006b:1). ROE can have a high value because of inadequate equity capital. By splitting the ROE into components, this problem can be solved (Sinkey, 1992:271). Increasing financial leverage means that the firm uses more debt financing relative to equity financing. A higher proportion of debt in the firm's capital structure leads to higher ROE. Increased debt will make a positive contribution to a firm's ROE only if the firms ROA exceed the interest rate on the debt (Anon., 2006b: 1). Sinkey (1992:271) divided ROE in the following three components:

t ~bJnt ltsif*s\w\n \ f C/if//5C l r Tn+nl Awn + v

ROE =

Net _ Income

Sales

Sales |

Total Assets

Total _ Assets

Average _ Equity

The equity multiplier (EM) equals total assets divided by total equity. EM represents a risk measure, because it reflects how many assets can go into default before a bank becomes insolvent (Koch & MacDonald, 2003:113).

ROE = ROAxEM (2.4)

According to Koch and MacDonald (2003:114), the du Font analysis can also be modified as follows:

J K M - - ^ - ™ - - ^ - ^ (2.5)

aTA aTA aTA aTA

thus:

ROA = AU-ER-TAX (2.6)

where:

Nl = Net Income TR = Total Revenue

EXP = Total operating expenses aTA = Average Total Assets AU = Total revenue divided by aTA

ER = Total operating expenses divided by aTA TAX = Applicable income taxes divided by aTA

TR consists of the net sales plus the sum of interest income, non-interest income and securities gains and losses. EXP equals the sum of interest expense, non-interest expense and provisions for loan and lease losses (Koch & MacDonald, 2003:114). ER represents the expense ratio, which indicates the efficiency of a bank to control expenses. AU represents asset utilization, which indicates the bank's ability to generate revenues (Sinkey, 1992:271).

The drawback of accounting ROE and ROA measures are that they do not include any risk adjustments (Bessis, 2005:10). They can vary substantially and will be highly dependent on the

industry, this is why it is best to compare them against ROE and ROA values of a similar company (Anon., 2006a: 1). The concept of risk-adjusted performance measures was developed because of the drawback of ROE and ROA and is the reason for moving to 'economic values', 'mark-to-market' or 'mark-to-model' values, because these are both risk- and revenue-adjusted (Bessis, 2005:10). ROA and ROE, that is traditional profitability measures, are beset with problems of allocating assets, equity and net income when applied at branch level (Smith & Schweikart, 1992:54).

2.5.1.3. Other profitability measurements

• Net interest margin (NIM): It is a summary measure of the net interest return on

income-producing assets (Koch & MacDonald, 2003:117):

Net _ Interest _ income

Average _ earnings _ assets

• Spread: It is a measure of the rate spread or funding differential (Koch & MacDonald, 2003:117):

/ J + , T \

Spread(SPRD) =

Interest Income

Average _ earnings _ assets

f r „ , , cv.„ A

Interest _ Expense

j y Average _ Interestbearing _ liabilties j

NIM and spread evaluates a bank's ability to manage interest rate risk (Koch & MacDonald, 2003:117).

• The burden ratio: It measures the amount of non-interest expense covered by fees, service charges, securities gains and other income as a fraction of average total assets (Koch & MacDonald, 2003:117):

„ , ,. [Nonlnterest Expense - Nonlnterest Income)

Burden ratio = ^ = — - = '- (2.9)

Average _ Total _ Assets

• The efficiency ratio: It measures a bank's ability to control non-interest expense relative to adjusted operating income (Koch & MacDonald, 2003:117):

^rr. . . Nonlnterest Expense

Efficiency _ ratio = =— (2.10)

Net Interest Income + Nonlnterest Income

Using profitability based ratios as the sole indices of branch performance will provide the decision maker with inadequate information (Eccles, 1991:132). Davenport and Sherman (1987:36) also stated that by concentrating on branch profit as a measure of performance can lead to equitable ways to allocate revenues and expenses.

2.5.2. Other financial measurements 2.5.2.1. Liquidity ratios

• Liquidity ratios refer to an organization's ability to meet obligations and if necessary, quickly convert assets into cash. Liquidity ratios contain the following ratios (Dennis, 2006:62):

• The current ratio that is equal to current assets divided by current liabilities. It is the standard measure of an organization's financial health. It also indicates if a customer is able to meet its current obligations.

• The quick ratio that is equal to current assets (less inventories) divided by current liabilities (Dennis, 2006:62).

2.5.2.2. Leverage ratios

Leverage ratios measure the contribution of stockholders and creditors and indicates to what extent the organization is reliant on debt financing. It also shows to what extent debt is used in

an organization's capital structure. Leverage ratios contain the following ratios (Dennis, 2006:62):

• The debt to equity ratio, which is equal to total liabilities divided by total equity. This ratio indicates the relationship between debt and the organization's capital (equity).

• The interest coverage ratio, which refers to earnings, before interest, taxes, depreciation and amortization, divided by interest expense. This ratio indicates the portion of debt interest covered by cash flow (Dennis, 2006:62).

2.5.2.3. Profitability ratios

Profitability ratios refer to the organization's ability to generate revenue, while encountering the cost of production. Profitability ratios contain the following ratios (Dennis, 2006:62):

• The gross profit margin that equals gross profit divided by total sales, where gross profit equals net sales minus cost of goods sold. This ratio indicates how efficient the use of materials and labour are in the production process. It also shows the percentage of net sales remaining after accounting for the cost of goods sold.

• The return on sales, which is net profit (net income after tax) divided by sales. This ratio indicates aftertax profit to sales (Dennis, 2006:62).

2.5.2.4. Efficiency ratios

Efficiency ratios measure the use of assets to generate sales and profits. Efficiency ratios contain the following ratios (Dennis, 2006:62):

• The payables turnover ratio, which is cost of sales divided by trade payables. This ratio indicates how quickly an organization can pay its bills.

• The inventory turnover ratio, which is cost of goods sold divided by average inventory. • The asset turnover ratio, which is net sales divided by average total assets. This

measure indicates how efficiently a firm utilizes its assets. A high ratio implies that the firm is using its assets efficiently to generate profits (Dennis, 2006:62).

• The account turnover ratio, which is total net sales divided by accounts receivable. The higher the ratio the faster the organization is collecting its receivables.

• The account receivable collection period, which is 365 days divided by the account

receivable turnover. This ratio indicates how many days are needed to collect all the accounts receivable.

• Days payable ratio, which is the cost of goods sold divided by the accounts receivable turnover. This ratio shows how many days it takes to pay the accounts payable.

• Days inventory ratio, which is 365 days divided by the inventory turnover. This ratio identifies the average length of time, in days, it takes the inventory to turn over (Anon., 2007:4-5).

• Sales to net worth ratio, which is the total sales divided by net worth. This ratio indicates how many sales in dollars are generated with each dollar of investment.

• Sales to total assets ratio, which is the total sales divided by the total assets. This ratio indicates how efficiently the organization generates sales on each dollar of assets. • Debt coverage ratio, which is the net profit plus any non-cash expenses divided by the

principal on debt. This ratio indicates the organization's ability to satisfy its debt obligations and its ability to take on more debt (Anon., 2007:5-6).

O'Donnell and Van der Westhuizen (2002:485-486) stated that the following ratios can also be used to measure bank performance: the ratio of bad debts to assets; the ratio of staff costs to assets plus liabilities; total costs per employee; the ratio of non-interest income to interest income; liquidity and credit risk associated with loan portfolio ratios; rates in deposit growth; net interest income (Nil) and net interest margin (NIM).

Several other financial ratios have also been used to detect the possibility of a financial crisis in bank companies. The popular ratios include:

Net Worth

Total Assets

Total __ Operating _ Income

Total _ Assets

Net Income

Total __ Assets

Current Assets

Current _ Liabilities

A problem with these ratios is that a bank may perform differently in different ratios (Kao & Liu, 2004:2355-2356). Beaver et al. (1970:660-661) stated that the following ratios should be considered in measuring bank performance:

• Dividend payout. Sometimes it is stated that firms with low payout ratios, like cash dividends available for common stockholders, are more risky.

• Growth, which can be defined as the growth in total assets, where the growth rate can be the result of one or all of the following:

o When the expected earnings rate on new asset acquisitions is greater than the cost of capital.

o When the ex post rate of return exceeds the expected return for several periods. o When a higher than average proportion of earnings is the result of a payout

policy.

• Leverage. The leverage ratio can be defined as the total senior securities plus the current liabilities divided by total assets. This ratio can be used as a measure of the risk induced by the capital structure.

• Liquidity. It is argued that current assets have a less volatile return than non-current assets.

• Asset size. Larger firms are less risky than smaller firms, in terms of default risk, because larger firms have a larger asset base.

• Variability in earnings. This measurement can be defined as the income available to

common stockholders in relation to the market value of common stock outstanding.

The Qatar Central Bank (2006:1-2) used some of the following ratios in their bank performance evaluations:

• Capital adequacy

Re gulatory _ Tier! _ Capital

o

Total _ Assets

• Asset quality standards

Non - Perfor min g __ Loans

Total _ Loans

Loans Provisions

o =

Total _ Loans

Net Profit

Average _ Shareholders' _ Equity

Liquidity standards

Liquid Assets

o

Total _ Assets

Liquid _ Assets

o

Liquid _ Liabilities

Loans to Pr ivate __ Sector

Pr ivate _ Sector _ Deposits

Loans to Private Sector

o

Domestic __ Credits

(Total _ Deposits + Shareholders' _ Equity)

o Growth Rate of Total Assets

o Growth Rate of Total Customers' Deposits o Growth Rate of Total Credit Facilities o Growth Rate of Financial Assets' Portfolio

Since most other financial ratios in bank performance analysis require that at least one of these components is calculated, they will also suffer from similar problems (Avkiran, 1997:224). Clark (1997:25) stated that financial measures are backward-looking and do not reflect the long-term and future consequences of managerial actions. Using financial indicators as the sole measure for incentive purposes may encourage managerial focus on the short-term and may distort the decision making process (Kaplan & Norton, 1996a:75).

The nature of financial ratios based on internal historical data means that they are of limited use for strategic decision making and planning. Ratios are not an effective means of representing the many facets of performance (Avkiran, 1997:224). Financial ratios used at the bank level are particularly difficult to generate at the branch level. The reason for this is due to problems associated with allocation of indirect variable and fixed costs, including capital revenues and expenses (Avkiran, 1997:224). Davenport and Sherman (1987:35) stated that: 'Ratios cannot capture the interplay among multiple resources and outputs'. Chelst et al. (1988:7) also stated that the traditional branch performance measures have the following disadvantages:

• Different types of deposits are treated equally, ignoring varying profit margins. • Revenues from loans are ignored.

Avkiran (1997:225) stated that using only accounting ratios in analyzing performance have the following problems:

• Accounting ratios give little indication as to the specific reasons for good or bad performance.

• What types of decisions enabled the organization to achieve a good performance.

In some cases the organization's decision makers can use non-financial measures to give them insight into the organization's operations. The next section will briefly discuss some non-financial measures.

2.6. Non-financial measures

According to Kaplan and Norton (2001a:11) more companies have placed greater emphasis on the use of non-financial measures. These measures include customer satisfaction, innovation measures, on-time delivery, market share, product/service quality and productivity. The use of non-financial measures has increased the track performance of organizations in the past (Kaplan & Norton, 2001a:4-7). The following reasons are given for the increased use of non-financial measures:

• Non-financial measures are believed to be better indicators of managerial effort (Johnson etal., 1995:705).

• Non-financial measures are better predictors of long-term performance and help managers on their long-term decisions and actions (Johnson & Kaplan, 1987:259).

• Non-financial measures deal with causes and not with effects (Johnson & Kaplan, 1987:256-257).

• Non-financial measures are also an important source of information on firm failure (Kaplan & Norton, 1996a:85).

• Non-financial measures are less prone to manipulation (Singleton-Green, 1993:52). • Non-financial measures can be used to measure the changing technological

environment and the key leverage capabilities in order to achieve the necessary competitive advantages (Eccles, 1991:133-134).

Berry et al. (2005:93-94) stated that no one single measure provides consistent evidence of the correlation between all stakeholders' satisfaction and firm performance. By using multiple measures of performance, measuring both financial and non-financial measures, led to the development of the following approaches:

• Benchmarking

• Total Quality Management (TQM)

• The European Foundation for Quality Management (EFQM) • The Balanced Scorecard (BSC)

Only the BSC will be discussed in the following section, because according to Kaplan and Norton (1992:79) it is the most favourite and most common approach used. Kaplan and Norton (2001 a:3) also stated that the BSC has the ability to translate the organization's mission and strategy into a comprehensive set of performance measures.

2.7. The Balanced Scorecard (BSC)

The BSC adds operation components to the traditional financial measures (Kaplan & Norton, 1992:71). The BSC enables management to link performance measures to their vision and strategy. It helps to channel the organization's resources, abilities and energy to achieve long-term goals (Kaplan & Norton, 1996:55-56). The BSC also allows the organization to pinpoint its strategic objectives by using the following four pillars (Kaplan & Norton, 1992:72):

2.7.1. Financial perspective

Financial measures have the ability to measure the economic consequences of actions already taken. The financial measures focus on profitability-related measures, such as ROE and return on sales (Kaplan & Norton, 1992:77).

2.7.2. Customer perspective

This measure includes overall indicators such as customer needs, the defect level of incoming and on-time delivery of required products and services (Kaplan & Norton, 1992:73). These measures can be calculated with the information gathered by customer surveys, repeated sales from customers and from customer profitability. The customer perspective is crucial, because it helps the organization to connect its internal processes to improved outcomes with its customers (Kaplan & Norton, 2001 b:93; 1996:58-61).

2.7.3. Internal business process measures

These measures are based on the objective of most efficiently produced products and services that meet customer needs. For example, conversion rate, on-time delivery from suppliers, cost of non-conformance and lead-time reduction (Kaplan & Norton, 1996:62-63).

2.7.4. Learning and growth measures

This pillar is all about developing the capabilities and processes needed for the future. This can include the speed of transactions or the number of people involved in a transaction (Kaplan & Norton, 1996a:84-85).

Chang and Chow (1999:395,410) stated that the BSC is most effective when used to drive organizational change and used in focusing on continuous improvement efforts. However, the BSC has some of the following limitations:

• BSC will be useless if the included non-financial measures are not linked to the firm's strategic objectives (Kaplan & Norton, 1996:55-56, 77).

• The BSC is weak if too many performance indicators are included (Gering & Rosmarin, 2002:18-19).

• The BSC is ineffective if it becomes a balanced brainstorm or a grab-bag of ideas to satisfy each strategic objective (Gering & Venkatraman, 2000:17).

• The BSC also fails when it focuses on trying to balance stakeholder conflict or when it acts as a management scorecard (Gering & Mhtambo, 2000:19).

• BSC does not take into account competitor actions, developments in technology, unexpected events and fails to establish a basis of continuous improvements (Norreklit, 2000:78).

Gering and Rosmarin (2002:19) however stated that some of these limitations can be overcome by following some of these recommendations:

• To use BSC as a centralized control.

• Do not try and balance the scorecard with financial and non-financial measures. • Do not use the BSC as a direct incentive system.

• Allow middle management to participate and contribute in selecting the appropriate technology with appropriate cost.

The problem in bank branches is to construct the balance sheets. The reason is that it is difficult to determine the amount of equity that must be assigned to each branch (Koch & MacDonald, 2003:175). Banks either did not allocate capital or allocated it on an undifferentiated basis according to the amount of assets employed, until the 1980s (Kimball, 1997:30). This led to the allocation of capital on a risk-adjusted basis and was the result of regulatory initiatives. The

failure to allocate capital based on the amount of risk involved can lead to serious performance measurement errors (Kimball, 1997:30). The expected proftit of each branch depends on the branch's preference towards risk, in other words, profit depends on risk (Pindyck & Rubinfeld, 2005:159-161). The drawback of accounting ROE and ROA measures are that they do not include any risk adjustments (Bessis, 2005:10).

In the next section (section 2.8) risk in bank performance evaluation is discussed, followed by the meaning of risk-adjusted measures, (section 2.9).

2.8. Risk-adjusted measures

Beaver et al. (1970:679) established the association between accounting-determined and determined measures of risk. A relationship was found between the measure of market-determined risk and accounting risk measures, for example, dividend payout and financial leverage. Louge and Merville (1972:38-44) investigated the relationship between the measure of market-determined risk and nine-book-value-based financial ratios. They found that only return on assets, asset size and financial leverage was statistically significant and explained the variability in systematic risk. Gonedes (1973:435-436) found evidence on the relationship between the information content of accounting income and market-determined risk measures. It suggested that accounting income did reflect a statistically significant amount of information imbedded in market prices of securities. Jahankhani and Lynge (1980:169,172-175) studied the association between market-determined and accounting-determined measures of risk in the banking industry. Total risk as the dependent variable and seven other variables were used. These variables included dividend payout, variation of earnings per share, loan to deposit ratio, loan loss experience and liquidity ratio. By using total risk as the dependent variable all accounting risk measures except loan to deposits ratio were statistically significant.

Mansur et al. (1993:1505-1506) used the following financial ratios, which are explained as follows:

• Stockholders' equity to total deposits ratio indirectly measures the effect of leverage on systematic and total risk.

• Total loans to total deposits ratio, which is positively related to both systematic and total risk.

• Net income to total assets ratio measures the bank's accrual income which reports all the managerial decisions and market events on the bank's profits.

• Total loans to total assets ratio, which is positively related to systematic and total risk. • Total loan loss reserve to total loans ratio, which is positively related to both

measures of risk.

• Cash and due from banks to total assets ratio, which measures the liquidity position of a bank.

• Coefficient of variation of deposits, which is positively related to market-determined risk.

Mansur et al. (1993:1506-1508) failed to explain the variability of market-determined risk and stated that the association between accounting- and market-determined risk measures have weakened over time.

Sinkey (1992:268-270) examined financial statements in an effort to analyze returns and risk. The main ratios reviewed were ROE, ROA, equity multiplier (EM), interest sensitivity ratios, liquidity ratios and the equity capital ratio. Managerial decisions are occupied with the balancing of returns against risk in reaching desired outcomes. To measure bank performance return and risk dimensions should therefore be examined (Avkiran, 2006:280). Returns can be measured by the ROE ratio, which can be decomposed into the equity multiplier (EM) and ROA. EM is the